Розробка системи розпізнавання рукописних символів нейронною мережею

Abstract

This diploma work developed software for handwriting symbol recognition using neural network. Were the basic pattern recognition methods defined their advantages and disadvantages, and based on this choice was made in favor of artificial neural network based on multilayer perceptron. It was the structure, mathematical model and algorithm chosen network, which served as the basis for the development of software algorithm. It was determined optimum potential user environment software, algorithm software and end up on the basis of the research was built character recognition software. The software is written in the programming language C#, software is characterized by clarity and ease of interface, high accuracy data processing, which provides basic user requirements for the organization recognition process of handwritten letters.